-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
LOSS can not convergence #1
Comments
Hi, thanks for your interests. I am wondering if Also, could you share the loss curve for the task of texture synthesis using your new forward function? This might be easier for debugging. |
Sorry, I don‘t understand why you mentioned batch_size = 1, slicing_torch(Ics_feats[0], style_feats[0]) where Ics_feats[0] is the output feature map of a layer of VGG after the stylized image, my style Loss finally dropped to about 50 and it couldn't go down anymore. |
No problem, yes the loss should not be that large. From my understanding, you don't need to call The issue may come from your new |
Hello, thank you very much for writing sliced_wasserstein_loss as a pytorch version. I tried to migrate it to my style transfer network and replace the original Gram matrix used to describe style features. Although the code runs successfully, Loss does I have not been able to converge. I have made some changes to the code. I don’t know if my understanding of the code is wrong. I hope you can help me. Thank you very much. The following are my changes:
Call this LOSS:
The text was updated successfully, but these errors were encountered: